Validation Of Clinical Prediction Rules For Bleeding In Patients With A Left Ventricular Assist Device.

Left ventricular assist devices (LVADs) reduce mortality and morbidity in patients with end-stage heart failure. Patients with LVADs require warfarin therapy, and bleeding occurs and complicates therapy in nearly half of these recipients. Predicting which patients will bleed after LVAD implant is challenging, and prediction rules for major bleeding have not been rigorously studied in this population. We aimed to validate previously described clinical prediction rules for bleeding in an LVAD cohort. This was a single-center, retrospective cohort study of patients who received an LVAD at Barnes-Jewish Hospital. Data used to tabulate risk scores and adjudicate outcomes were collected through chart review. The primary clinical endpoint was time to major gastrointestinal bleed or intracranial hemorrhage after implant. The secondary endpoint was time to any major bleed after hospital discharge. A total of 418 patients received an LVAD (135 HeartMate II, 125 HeartMate 3, 158 HVAD) between November 2009-January 2019. The primary endpoint occurred in 168 (40.1%) patients. C-statistics for the primary endpoint ranged 0.53-0.58 (P<0.05 for HAS-BLED, HEMORR 2 HAGES, ATRIA, Outpatient Bleeding Risk Index, and VTE-BLEED). The secondary endpoint occurred in 167 (40.0%) patients, and C-statistics ranged 0.52-0.58 (P<0.05 for HAS-BLED, Outpatient Bleeding Risk Index, and VTE-BLEED). After modifying the age threshold in each prediction rule to 59 years (the median age of the cohort), C-statistics ranged 0.55-0.61 for the primary and secondary endpoints. In a subgroup analysis of HeartMate 3 patients, prediction rules performed similarly. Existing clinical prediction rules for major bleeding, previously derived in other anticoagulated populations, had mediocre discrimination in an LVAD cohort. [ABSTRACT FROM AUTHOR]